Michael P. Manti

• Led team of seven data scientists in Collateral Modeling and Analytics (CMA) responsible for development of Home Value Explorer (HVE), Freddie Mac's automated valuation model (AVM), which predicts sale prices of individual homes.
• Oversaw enhancement of ongoing monitoring (OGM) of HVE to include interpretability and model drift and improve benchmarking.
• Coordinated multi-team effort to develop faster, more accurate models to serve as challengers to HVE.

• Led Model Liaison team in Modeling, Econometrics, Data Science, and Analytics (MEDA) department of Freddie Mac’s Single-Family business division. Supervised three analysts.
• Acted as intermediary between model owners in MEDA and three oversight groups: Enterprise Model Risk (EMR), Internal Audit (IA), and Freddie Mac’s regulator, Federal Housing Finance Agency (FHFA).
• Facilitated targeted examinations and ongoing monitoring of MEDA’s models and model processes by FHFA. FHFA conducted four such examinations per year on average.
• Facilitated audits of MEDA’s models and model processes by IA. IA conducted three such audits per year on average.
• With Regulatory Affairs, coordinated MEDA’s responses to requests from FHFA. Enhanced completeness of responses and shorten response timelines.
• Collaborated with model owners on design of agreed-upon actions (AUAs) to remediate findings from IA and matters requiring attention (MRAs) from FHFA.
• Established robust processes for managing examinations, audits, and issues by developing and maintaining playbooks, checklists, reports, and calendars.
• Advised model owners on application of enterprise-wide standards and divisional procedures to their models.

• Supervised team of four data scientists in Collateral Modeling and Analytics (CMA) department of Freddie Mac’s Single-Family business division. Team was among first to use tools for big data (e.g., Apache Spark) and automated machine learning (AutoML) at Freddie Mac.
• Researched, developed, tested, and documented rules and model for automating several collateral underwriting decisions in two of Single-Family’s business offerings.
• Contributed to design of application programming interface (API) to provide Single-Family's customers with early read on eligibility for one of its offerings.

• Validated risk rating, probability of default (PD), loss given default (LGD), and exposure at default (EAD) models for Capital One's commercial portfolio.
• Supervised two quantitative analysts.
• Conducted annual reviews for 17 models.
• Effectively challenged model developers to improve performance monitoring and development methodology for several models.
• Taught course on technical writing to model developers and validators. Updated course materials to be consistent with new model policy. Course was well received by students.

• Managed research, development, testing, documentation, and execution of four of State Street's models for estimating capital.
• Supervised four quantitative analysts.

• Managed research, development, testing, documentation, and execution of State Street's Operational Risk Capital Model.
• Supervised two quantitative analysts.
• Presented methods and results to technical and non-technical audiences, including State Street executives and Federal Reserve examiners.

Updated quantitative models underlying State Street's credit rating tools for evaluating bank and corporate counterparties.

• Reviewed models used to manage credit risk and produce financial statements.
• Supported credit risk oversight with quantitative analysis.
• Supervised two quantitative analysts.

• Managed team of four quantitative analysts who researched problems with, wrote requirements for, and tested implementation of a complex stress-test model.
• Collaborated with Freddie Mac's safety and soundness regulator on requirements, testing, and deployment of software that determined regulatory risk-based capital requirements for Freddie Mac.